NWS FROST DEPTH OBSERVATION WITH LIQUID-IN
PROBES PERFORMANCE: TWO-YEAR REVIEW
F. Adnan Akyüz
North Dakota State University
Mark Ewens
National Weather Service
Grand Forks Weather Forecast Office
Radu Carcoana and Barbara Mullins
North Dakota State University
2008
Journal of Service Climatology
Volume 2, Number 2, Pages 1-10
A Refereed Journal of the American Association of State Climatologists
NWS Frost Depth Observation with Liquid-In Probes Performance:
Two-Year Review
F. Adnan Akyüz
North Dakota State University
Mark Ewens
National Weather Service
Grand Forks Weather Forecast Office
Radu Carcoana and Barbara Mullins
North Dakota State University
High Plains Regional Climate Center
School of Natural Resources
University of Nebraska
Corresponding Author: Adnan Akyüz . School of Natural Resource Sciences, 106 Walster Hall, North
Dakota State University, Fargo, ND 58107, USA. Tel. 1-701-231-6577, [email protected].
(Non-technical summary is at the end of this paper)
Abstract
Performance of the liquid-in frost depth probe made in-house by the Grand Forks National
Weather Service (NWS) Weather Forecast Office (WFO) is compared against soil temperature
observations made by North Dakota Agricultural Weather Network (NDAWN) at the Fargo
location for a two-year period from 2006-2007 to 2007-2008 winter seasons. While the liquid-in
frost depth probe provided continuous frost depth observations, NDAWN soil temperature
observations had to be interpolated between measurement points to determine the depth where
the soil temperature was 0 °C (32 °F). In general, the trends of both observations matched almost
identically; however, the magnitude of the depths varied with liquid-in frost depth probe
consistently showing deeper frost depths than the NDAWN soil temperature observations.
Differences are discussed further in the following sections.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
1
1. The Observation Site
The North Dakota Agricultural Weather
Network (NDAWN) station number one (1) in
Fargo is located on 0.6 hec (1.5 acres) of flat,
grassy terrain just to the west of the North
Dakota State University (NDSU) campus and
less than 1.6 km (one mile) south of Fargo
Hector International Airport (46°53'50.44"N,
96°48' 43.65"W), NDSU Microclimate Research
Station (MRS). The observation area in NDSUMRS is 67 m (220 feet) by 98 m (320 feet). It is
protected with wire-fence to stop curious
passers-by tampering with the sensors. The
observation site is a host site to other research
instruments such as UV-B station number 27
(USDA, 2008), lysimeter (Casey et al., 2008),
and others (Smerdon et al., 2006; Smerdon et al.,
2004; Smerdon et al., 2003; Smerdon et al.,
2002; and Schmidt et al., 2001) . In addition, the
NDAWN station at this site interfaces with a
deep soil moisture profile that samples soil
temperature data at multiple depths from 1 cm to
11.7 m. Therefore, it was a natural choice for the
National Weather Service (NWS) frost depth
probe to be placed next to the NDAWN weather
station.
2. North Dakota Agricultural
Weather Network (NDAWN) Deep
Soil Temperature Profile
The NDAWN network is comprised of 70
automated weather stations across North Dakota,
western Minnesota, eastern Montana and
northern South Dakota. All weather stations in
the network records hourly basic weather
elements: air temperature, rainfall, wind speed
and direction, solar radiation, soil temperature
at 10 cm (4 inches) under bare and turf
conditions, and atmospheric pressure (in most
stations). In addition, the NDAWN station in
Fargo collects hourly deep soil temperatures at
multiple depths. The data have been collected
daily since 1992. Type T-thermocouples (Extension grade, copper/constantan, duplex, ANSI
type TX) are deployed at 23 different depths: 1,
5, 10, 20, 30, 40, 50, 60, 80, 100, 125, 150, 175,
200, 250, 270, 300, 370, 470, 570, 770, 970,
1170 cm below the ground level.
Figure 1 depicts the measurement points in
the profile. These thermocouples can easily be
purchased as spools and can be cut to any
desired length. Each thermocouple has two leads
Figure 1. NDAWN-Fargo Station Deep Soil Temperature Profile Measurement Points.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
2
with dissimilar metals (copper and constantan)
that respond to temperature changes as voltage
gradient. We then translated those voltage gradients into temperature values using a CR10X
datalogger (Campbell Scientific Inc.). Since
there were 23 sensors with 46 leads (two leads
for each sensor) would require more channels
than the datalogger could provide, we added
more channels using a 32-channel relay
multiplexer that allowed 32 additional
differential two-wire sensors at a time.
3. NWS Frost Depth Observations
The North Central River Forecast Center
(NCRFC) currently uses the Sacramento Soil
Moisture Accounting (SAC-SMA) model as part
of the Advanced Hydrologic Prediction System
(AHPS) suite. The SAC-SMA parameterizes the
depth of frozen ground in order to calculate the
runoff/absorption ratio during spring snow-melt
flood scenarios. In order to gain a better understanding of how the frost within the Red River
Valley of the North (RRVoN) actually behaves
during the spring snow-melt period, the Weather
Service (NWS) in Grand Forks was tasked with
the development and deployment of a frost
probe network.
The NWS Grand Forks frost probes were
developed based on the Brian Hahn (Service
Hydrologist, Milwaukee WI) model employed in
Wisconsin in 2004-2005. The Hahn model
employed a Tygon™ plastic tube 2.54 cm by
1.83 m (1 inch x 6 foot) in size filled with water
colored with Orcoacid™ fluoresciene dye,
placed in a 5.1 cm (2 inch) diameter PVC sleeve,
then buried. Orcoacid™ fluoresciene dye is an
orange-colored compound that turns green when
mixed with tap water. During the freezing
process, the frozen layer becomes clear, making
the determination of the frost depth easy.
The NCRFC and NWS Grand Forks selected
16 locations within the RRVoN that would
suitably represent the distribution of frost as
described by the SAC-SMA. Table 1 provides a
list of these locations. The Grand Forks variant
differs slightly from the Hahn version, in that we
used 1.00 cm by 1.52 m (3/8 inch x 5 foot)
lengths of Tygon™ tubing placed in a 1.6 cm
(5/8 inch) PVC housing. Increments of 2.54 cm
(1 inch) were drawn on the exterior of the
Tygon™ plastic tubing, and the tube filled to the
zero level with the water/Orcoacid™
fluoresciene dye mixture (Figure 2).
Operational deployment was performed
during September and October 2006. Of the 16
total, the original probe was installed at NWS
Grand Forks in October 2005 as an Operational
Test & Evaluation. Twelve probes were installed
during the fall of 2006, with two additional
probes
Table 1. NWS Frost Depth Probe Locations in the Red
River Valley of the North.
City/State
1. Grand Forks, ND NWS
2. Langdon, ND
3. Camp Norris, MN
4. Endmore 1NW, ND
5. Baldhill Dam, ND
6. Thief River Falls, MN
7. North Dakota State University/Fargo ND
8. Casselton Ag. Farm, ND
9. Lisbon, ND
10. Forman, ND
11. Halstad, MN
12. Orwell Dam, MN
13. Campbell, MN
14. Pembina, ND
15. Dalton 3S, MN
16. Devils Lake, ND
________________________________________________
installed during the summer of 2007. All probes
were co-located with official NWS Cooperative
Weather Observation sites. The Fargo probe was
installed in July 2006 and the first data were
recorded on December 1, 2006.
Figure 2. Tygon™ plastic tube filled with water/
Orcoacid™ fluoresciene dye mixture
Installation was performed by drilling a 5.1
cm (2 inch) diameter hole to a depth of 1.82 m
(6 feet). The PVC housing was lowered to a
depth of 1.52 m (5 feet; predetermined by
external reference marking), then the hole was
backfilled with crystalline silica, commonly
called “sandblasting sand”. The intent of the
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
3
Figure 3. 10 cm (4 inch) protective PVC cover and its lid.
crystalline silica was to offer stability in the core
and maintain proper conduction in the soil. The
probe was protected by covering the access with
a 10 cm (4 inch) PVC cover and lid (Figure 3).
There are several key assumptions that need
be stated as a point of clarification. First is that
the Orcoacid™ fluoresciene dye did not
sufficiently modify the specific or thermal
density of the water in the probe to affect the
freeze rate or natural conductive properties. In
this way, the depth measured was accurate. The
second critical assumption is that the
water/Orcoacid™ fluoresciene dye mix actually
froze at 0 °C. Although most frost data were
collected weekly on Monday mornings at the
same time as the normal cooperative observation
time, the Fargo location collected frost depth
measurements at noon. Then the data were
transmitted to the Grand Forks NWS office via
telephone. Most other observers typically used
the WxCoder II web based program (now
WxCoder III), or called NWS Grand Forks for
encoding and transmission. When requested by
the NCRFC, additional data could be collected
and transmitted.
4. Analysis
As seen in Figure 1, more soil temperature
measurement points were selected at depths
closer to the surface than deeper into the soil so
that the frost depth is captured in the closest
proximity to the surface. However, the
measurement depths were discrete variables, so
exact frost depth had to be interpolated between
the measurement points. For example, if the
frost depth was between the 10 and 20 cm levels
we used linear, polynomial and logarithmic
interpolation to estimate the actual frost depth.
All three methods came within 1% from each
other (i.e. 1 mm difference at 10 cm depth).
Therefore, we used linear interpolation for
simplicity if 0 °C occurred between two
measurement points to estimate the frost depth.
We also assumed that the ground would freeze
at 0 °C. The authors realized that the assumption
might not be applicable when comparing the
frost depths among different soils with different
water content. We also assumed that when we
measured 0 °C at a depth, the soil would freeze
the liquid in the NWS liquid-in frost depth
probe.
The ground normally freezes after air
temperatures fall below 0 °C continuously in
fall. Radiant and conductive heat exchange
occurs at the surface; therefore, there may be
some diurnal fluctuation near the surface.
However, these diurnal fluctuations decrease
rapidly with depth. Heat exchange within soil is
through conduction, which occurs very slowly.
Frost depth increases with time well into spring
in Fargo. When the air temperature reaches
above freezing in spring, heat exchange allows
ground to thaw at the surface first. At this time,
frost depth at the bottom of the frozen layer of
soil remains relatively unchanged. As the
surface layer of thawed soil deepens in spring, a
layer of frozen soil is trapped underneath.
Eventually this layer thaws in late spring. Hence,
there are two frost depths in the soil at any given
time: at the top and the bottom. In fall and
winter, the ground is frozen from surface to
bottom. Therefore the top depth is zero.
Both the top and the bottom frost depth
measurements have different applications. When
planting crops, for example, farmers may be
interested in top frost depths while the
construction industry may be more interested in
the bottom frost depths.
We analyzed both the top and the bottom
frost depths for a two-year period from fall 2006
to spring 2008. Since the top frost depth is zero
in fall, we only analyzed the top during spring
seasons. Our analysis was comprised of
comparing frost depths measured by the NWS
liquid-in frost depth probe with the frost depths
estimated by the NDAWN deep soil temperature
profile in Fargo, ND. In this study, estimation is
attributed to the NDAWN deep-soil temperature
profile because the data were not observed in a
spatially continuous setting. Actual frost depths
were interpolated linearly within the profile
using two measurement points on each side of
the estimated frost depth.
Table 2 is the dataset of two-year side-byside observations of frost depth made in NDSU
Microclimate Research Station, Fargo. In the
spring of 2008, the NWS frost depth probe
showed 5 cm of missing liquid. Therefore, a
missing value was entered if the actual top frost
depth was less than 5cm.
Figures 4 and 5 are the comparison graphs
between the NWS frost depth probe and the
NDAWN soil temperature profile in the 2006-07
and 2007-08 winter seasons respectively. They
both show the bottom frost depths.
Figures 6 and 7 are the comparison graphs
between the NWS frost depth probe and the
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
4
Table 2. A Two-year data of the NWS Frost Depth
measurement and the NDAWN estimation of frost
depth in Fargo, North Dakota (Top and Bottom
represent the top and the bottom of the frost layer
respectively)
Date
12/1/2006
2/1/2007
3/5/2007
3/12/2007
3/19/2007
3/26/2007
3/28/2007
3/30/2007
4/2/2007
4/9/2007
4/16/2007
4/20/2007
4/23/2007
11/26/2007
12/6/2007
12/18/2007
12/27/2007
1/2/2008
1/8/2008
1/14/2008
1/22/2008
1/28/2008
2/4/2008
2/12/2008
2/18/2008
2/25/2008
3/7/2008
3/17/2008
3/24/2008
3/31/2008
4/7/2008
4/9/2008
4/13/2008
4/18/2008
4/21/2008
4/23/2008
Top
(cm)
0.0
0.0
0.0
-5.1
-12.7
-22.9
-38.1
-30.5
-22.9
0.0
-12.7
-40.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
missing
missing
missing
missing
-7.6
-15.2
-35.6
-48.3
0.0
NWS
Bottom
(cm)
-12.7
-76.2
-76.2
-76.2
-76.2
-73.7
-71.1
-71.1
-71.1
-71.1
-68.6
-66.0
0.0
0.0
-12.7
-20.3
-20.3
-20.3
-20.3
-22.9
-38.1
-47.0
-49.5
-61.0
-66.0
-71.1
-76.2
-81.3
-81.3
-81.3
-81.3
-76.2
-73.7
-71.1
-71.1
0.0
Top
(cm)
0
0
0
0
0
-17.9
-26.3
-28.7
-29.4
0
-28.1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-9
-4
-8
-9
-29
-28
-61.5
-60
0
NDAWN
Bottom
(cm)
-14.5
-53.7
-61.7
-59.1
-60
-52.7
-49.9
-50.7
-52.5
-55.8
-49.1
0
0
0
-4.5
-6.6
-8
-14.2
-14
-24.1
-36.3
-43
-51.6
-55.8
-63.3
-65.4
-69.7
-64.4
-63.6
-62.2
-60.2
-63.2
-57.8
-61.5
-60
0
NDAWN soil temperature profile in 2007 and
2008 spring seasons respectively. They both
show the top frost depths. In 2008, the NWS
frost depth tube had 5 cm (2 inches) of loss in
liquid level. If the frost occurred between the
surface and 5 cm layer, there was no way of
knowing where the frost depth was based on
NWS frost depth probe, therefore, it was not
plotted.
5. Conclusions
a. Bottom frost depths
•
•
•
•
•
Both 2006-07 and 2007-08 winter season experiments indicated that the frost
depths measured with the NWS liquid-in
frost depth probes were deeper.
NWS frost depth probe showed less
variation with time suggesting time
delay in liquid responding the temperature change in the soil.
Differences between the NDAWN soil
temperature profile estimation and the
NWS frost depth probe were greater
with depth.
Maximum difference was 66 cm when
the NWS probe measured the bottom of
the frost as 66 cm below surface and the
NDAWN soil temperature profile
measured no temperatures at or below 0
°C on April 20, 2007. This should not be
misinterpreted. It is possible that a thin
layer of frozen soil was trapped between
the top and the bottom frost depths.
When the temperature is at or around 0
°C, small differences in temperature
measurements with different methods
can misleadly appear as big differences
in frost depth observations.
In 2006, the NWS observation was
started too late. When the first observation was recorded on December 1,
2006, the bottom of the frost depth was
already 13 cm. In 2007, the first day of
the frost was successfully captured.
b. Top Frost Depths:
• Both measurement methods were highly
depended on air temperature since
radiant heat exchange between the air
and the soil takes place at the surface.
• Seasonal variations of the top frost
depths were greater than those of the
bottom frost depths.
• In 2008, the NDAWN estimation
showed deeper top frost depths than the
NWS measurements while in 2007
there was no distinct pattern.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
5
Figure 4. 2006-2007 Winter Season Bottom Frost Depth Comparison between the NWS Frost Probe (blue line) and the NDAWN
Deep Soil Temperature Profile (green line).
Figure 5. 2007-2008 Winter Season Bottom Frost Depth Comparison between the NWS Frost Probe (blue line) and the NDAWN
Deep Soil Temperature Profile (green line).
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
6
Figure 6. 2007 Spring Season Top Frost Depth Comparison between the NWS Frost Probe (blue line) and the NDAWN Deep
Soil Temperature Profile (green line).
Figure 7. 2008 Spring Season Top Frost Depth Comparison between the NWS Frost Probe (blue line) and the NDAWN Deep
Soil Temperature Profile (green line). NWS frost depth tube had 2.5” (5 cm) of loss in liquid level in 2008. Therefore the frost
depths above 5 cm (2 inches) were not plotted.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
7
• While the NDAWN estimation reached
above freezing conditions later than the
NWS measurements showed in 2007, it
was the opposite in 2008. In fact, when
the NDAWN estimation showed that
the top frost depth was 9 cm on April 7,
NWS frost depth probe was frozen from
0 to 81 cm on the same day. The
observer noticed that the probe lost 5
cm of liquid. We believe that if the
difference would have been less if the
liquid level was sufficient.
c. Recommendation to NWS:
•
•
•
•
•
•
When the liquid in the Tygon™ plastic
tube freezes, it expands. A permanent
plug at the bottom of the tube is
necessary so that the liquid does not
escape the tube. Based on personal
communication with the local NWS
office in Grand Forks, this will be
implemented in the next generation.
Thermal properties and specific density
of the water/dye mixture versus plain
water need to be defined. Effects of the
expansion of the liquid and the plastic
tube in the depth reading is unknown. If
the expansion is significant, it should be
taken into account when the depth
markings are labeled on the tube.
At the end of each summer season,
liquid level need to be checked so that
zero coincides with the surface. As seen
in Figure 2, the liquid level is reduced
more than 5 cm (2 inches). This could
be hampered by staffing issue within the
local NWS office. This practice could be
implemented as part of coop station visit
by the Data Administration Program
Manager.
More frequent readings are necessary.
There are big gaps between the reading
times when the frost depth changes.
Daily readings are necessary in spring
and fall to capture the gradual increase
and decrease in frost depth.
The freezing point of the liquid used in
the tube needs to be verified and
accurately defined in laboratory settings
outside of the NWS local office ability.
It also needs to match the native soil’s
thermal properties (Ochsner & Baker,
2008).
The 1.6 cm (5/8 inch) PVC housing is
separated from the soil with more than
2.5 cm thick crystalline silica for
stability purposes. However, its impact
on the native soil’s thermal property is
unknown. The authors recommend
drying the same soil that came
out of the hole and breaking in
smaller pieces to back fill the
hole (this would require
additional trips to the site, or
o using an auger smaller than 5
cm (2 inches) in diameter to
minimize the amount of space
that needs to be filled, or
o using bentonite clay instead to
maintain
stability
without
drastically changing the native
soil’s thermal properties. This
would work well in RRVoN
where soil is mostly clay
(NRCS, 2008). It may not work
well in soil with high sand
content. Bentonite clay is
commercially available and
could be purchased for the
Fargo site to do parallel studies.
The air gap between the 1.0 cm (3/8
inch) Tygon™ plastic tube containing
the liquid and the 1.6 cm (5/8 inch) PVC
plastic tube needs to be minimized to
eliminate convection in free air. Cold air
can easily sit at the bottom of the gap
causing NWS probes to overestimate the
bottom frost depth. Both in 2007 and
2008, the NWS consistently showed
deeper bottom frost depths than the
NDAWN estimation suggested. After
the achievement of a smaller gap, NWS
will have to employ tubing with zero
expansion to reduce the risk of jamming.
Finally, the authors of this paper realize
the potential benefits of this application
especially in flood forecasting, therefore
strongly encourage the NWS to test and
calibrate the probes in a laboratory to
address these issues before the national
implementation. Local NWS offices
deploying these practices may need to
contact their regional headquarters for
regional and federal assistance.
o
•
•
Acknowledgments
The authors would like to thank Dr. Laura
Overstreet, soil scientist in Soil Science
Department, North Dakota State University for
sharing her expertise on the RRVoN soil
properties.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
8
4. References
Casey, F., P. Oduor, H. Hakk, B. Larsen, and
T. Desutter, 2008. Transport of 17beta-estradiol
and testosterone in a field lysimeter . Soil
Science. 173 (7). 456-467.
NRCS. (n.d.). SSSC Soil Survey Labratory
Soil Characterization Database. Retrieved July
2008, from http://ssldata.nrcs.usda.gov/
Ochsner, E. T., & Baker, J. M., 2008. In-Situ
Monitoring of Soil Thermal Properties and Heat
Flux during Freezing and Thawing. Soil Science
Society of America, 72:1025-1032. doi:
10.2136/sssaj2007.0283.
Smerdon, J. E., H. N. Pollack, V. Cermak, J.
W. Enz, M. Kresl, J. Safanda, and J. F.
Wehmiller, 2006. Dalily, Seasonal and Annual
Relationship between air and subsurface
temperatures. J. Geophys. Res., 111, D07101,
doi: 10.1029/2004JD005578, 2006.
Smerdon, J. E., H. N. Pollack, V. Cermak, J.
W. Enz, M. Kresl, J. Safanda, and J. F.
Wehmiller, 2004. Air-ground Temperature
Coupling and subsurface Propagation of Annual
Temperature Signals. J. Geophys. Res., 109,
D21107, doi: 10.1029/2004JD005056.
Smerdon, J. E., H. N. Pollack, J. W. Enz,
and M. J. Lewis, 2003. Conduction-dominated
Heat Transport of the Annual Temperature
Signal in Soil. J. Geophys. Res., 108(B9), 2431,
doi: 10.1029/2002JB002351.
Smerdon, J. E., M. J. Lewis, H. N. Pollack,
and J. W. Enz, 2002. Spectral Analysis of Air
and Ground temperatures at Fargo, North
Dakota: Conduction Dominated Propagation of
the Annual Frequency Signal. EOS Trans. AGU,
83 (47), Fall Meet. Suppl. Abstract PP51A0304.
Schmidt, William L., W. D. Gosnold, and J.
W. Enz, 2001. A decade of Air-Ground Temperature Exchange from Fargo, North Dakota.
Global and Planetary Change. 29:311-325.
US Department of Agriculture. (n.d.).
http://uvb.nrel.colostate.edu/UVB/.
Retrieved July 2008, from
http://uvb.nrel.colostate.edu/UVB/
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
9
Non-Technical Service Summary for Akyüz, F.A. et al., 2008. NWS Frost Depth
Observation with Liquid-In Probes Performance: Two-Year Review.
The Climate Service Question:
How did the NWS frost depth probes perform during the first two years of their deployment in North
Dakota and are they ready for national deployment?
Background Information:
The North Central River Forecast Center (NCRFC) currently uses the Sacramento Soil Moisture
Accounting (SAC-SMA) model as part of the Advanced Hydrologic Prediction System (AHPS) suite.
The SAC-SMA parameterizes the depth of frozen ground in order to calculate the runoff/absorption ratio
during spring snow-melt flood scenarios. In order to gain a better understanding of how the frost within
the Red River Valley of the North (RRVoN) actually behaves during the spring snow-melt period, the
National Weather Service (NWS) in Grand Forks was tasked with the development and deployment of a
frost probe network.
Summary of the Research Methodology:
Performance of the liquid-in frost depth probe made in-house by the Grand Forks National Weather
Service (NWS) Weather Forecast Office (WFO) is compared against soil temperature observations made
by the North Dakota Agricultural Weather Network (NDAWN) at Fargo location for a two-year period
from 2006-2007 to 2007-2008 winter seasons. While the liquid-in frost depth probe provided continuous
frost depth observation, NDAWN soil temperature observation had to be interpolated between
measurement points to determine the depth where the soil temperature was 0°C (32°F).
Discussion:
We analyzed the top and the bottom frost depths for two-year period from 2007 to 2008. Since the top
frost depth is zero in fall, we only analyzed the top during spring seasons. Our analysis is comprised of
comparing frost depths measured by the NWS liquid-in frost depth probe with the frost depths estimated
by the NDAWN deep soil temperature profile in Fargo, ND. In this study, estimation is attributed to the
NDAWN deep-soil temperature profile because the data were not observed in a spatially continuous
setting. Actual frost depths were interpolated linearly within the profile using two measurement points on
each side of the estimated frost depth.
Both sets of data, from NWS frost depth probe and the NDAWN data, showed similar trends. However,
the magnitude of depths was dissimilar. In general, the NWS frost depth probse suggested deeper bottom
frost depths than that of suggested by the NDAWN data.
The main body of the paper includes recommendations to NWS to minimize errors associated with
measuring the level at which soil temperature is 0°C (32°F). However, the recommendations are
applicable to whoever is willing to measure this depth for varying purposes similar to or other than the
NWS objectives. Probe specifications (size, material, properties, etc.) and pictures are provided in the
paper. The probe can be built in-house easily with commercially available materials. It is relatively
inexpensive compared to commercially available multiple soil temperature probes located at multiple
depths under soil surface interfaced to a datalogger for temporally continuous scale. However, one of the
most important assets that the NWS frost depth probe possesses is that it provides spatially continuous
measurement scales.
Akyüz, F. A. et al., 2008. Journal of Service Climatology. www.journalofserviceclimatology.org
10